Comparative analysis of monoclonal antibody N-glycosylation using stable isotope labelling and UPLC-fluorescence-MSby Silvia Millán Martín, Cédric Delporte, Amy Farrell, Natalia Navas Iglesias, Niaobh McLoughlin, Jonathan Bones

The Analyst

About

Text

Analyst

COMMUNICATION

Cite this: Analyst, 2015, 140, 1442

Received 19th December 2014,

Accepted 15th January 2015

DOI: 10.1039/c4an02345e www.rsc.org/analyst

Comparative analysis of monoclonal antibody

N-glycosylation using stable isotope labelling and UPLC-fluorescence-MS†

Silvia Millán Martín,a Cédric Delporte,a,c Amy Farrell,a Natalia Navas Iglesias,b

Niaobh McLoughlina and Jonathan Bones*a

A twoplex method using 12C6 and 13C6 stable isotope analogues (Δmass = 6 Da) of 2-aminobenzoic acid (2-AA) is described for quantitative analysis of N-glycans present on monoclonal antibodies and other glycoproteins using ultra performance liquid chromatography with sequential fluorescence and accurate mass tandem quadrupole time of flight (QToF) mass spectrometric detection.

N-Glycans present at the asparagine 297 residue in the CH2 domain of the Fc region of IgG play a crucial role in modulating the structural stability and functional activity relationship of the antibody.1,2 Quantitative structural characterisation of

N-glycans present on antibody therapeutics, such as monoclonal antibodies (mAbs) and Fc fusion proteins, is required under the regulatory guidelines. Characterisation of the oligosaccharides present is normally performed using liquid phase separation techniques with optical detection following enzymatic liberation of the N-glycans from the antibody and derivatisation with a fluorophore to increase detection sensitivity.3,4

Despite recent advances in separation chemistries for oligosaccharides,5,6 peak area based quantitation remains a subjective process due to integration issues for distorted peaks containing poorly or partially resolved oligosaccharides, or due to the presence of multiple glycans within a single chromatographic peak.7 Such effects can result in either under reporting of a particular N-glycan or create difficulties in the alignment of data for subsequent statistical evaluation.

Relative or absolute quantitation, for the elucidation of differential expression of proteins, using stable isotope labelling technology has become widely used in proteomics.8 Application of stable isotope methods for quantitative glycomics has also attracted attention with published reports focusing upon heavy isotope incorporation during permethylation of oligosaccharides prior to MALDI-MS analysis.9–11 A limitation to this approach is that isotope incorporation and the associated light–heavy mass difference varies with the number of methylation sites available on the oligosaccharide. This limitation has been overcome through the labelling of oligosaccharide samples for comparison with either 13CH3I or 12CH2DI, respectively, thereby introducing a mass difference of 2.922 mDa per methylation site, however high resolution

FT-ICR-MS was required to distinguish between the resulting isotopic pairs.12,13 Stable isotope incorporation using reductive amination has also been reported prior to MALDI-MS analysis using 13C6 or deuterated analogues of commonly used reagents for high sensitivity fluorescence detection.14–21

Specifically designed hydrophobic hydrazide reactive reagents have been reported that facilitate MS based quantitation based upon 13C6 incorporation and also increase the hydrophobicity of the glycoconjugate, thereby, increasing its ionisation efficiency.22,23 Metabolic generation of isotopically enriched aminosugars and their products resulting from 15N glutamine feeding of mammalian cells, in an analogous manner to heavy lysine or arginine incorporation through SILAC, has also been described.24,25 MS/MS isobaric tagging methods were also reported for quantitative glycomics, however concerns exist regarding the attainment of a balance during fragmentation to ensure complete reporter ion release while avoiding over fragmentation of the oligosaccharide being characterised.26–29 A limitation of many of the cited methods is their use with either direct infusion ESI-MS or MALDI-MS, thereby limiting the possibility to gain quantitative information for positional or linkage oligosaccharide isomers present in the N-glycans pools under study due to the lack of an analytical separation.

Here, the use of 12C6- and 13C6 2-aminobenzoic acid (2-AA) differential labelling of N-glycans followed by ultra performance †Electronic supplementary information (ESI) available: Detailed methods and experimental procedures and analytical data for the forced degradation studies as mentioned in the main text. See DOI: 10.1039/c4an02345e aCharacterisation and Comparability Laboratory, NIBRT – The National Institute for

Bioprocessing Research and Training, Foster Avenue, Mount Merrion, Blackrock, Co.,

Dublin, Ireland. E-mail: jonathan.bones@nibrt.ie; Tel: +353 1215 8105 bDepartment of Analytical Chemistry, Faculty of Science, Biomedical Research

Institute, University of Granada, Avenue Fuentenueva S/N, 18071 Granada, Spain cLaboratory of Pharmaceutical Chemistry & Analytical Platform of the Faculty of

Pharmacy, Université Libre de Bruxelles, Bld. du Triomphe, Campus Plaine,

CP 205/05, B-1050 Brussels, Belgium 1442 | Analyst, 2015, 140, 1442–1447 This journal is © The Royal Society of Chemistry 2015

Pu bl ish ed o n 15

Ja nu ar y 20 15 . D ow nl oa de d by

K an sa s S ta te

U ni ve rs ity o n 25 /0 2/ 20 15 1 1: 16 :5 5.

View Article Online

View Journal | View Issue hydrophilic interaction liquid chromatography with online fluorescence and tandem mass spectrometry using a quadrupole time of flight instrument (UPLC-Flr-QToF-MS/MS) is described for the comparative analysis of the glycosylation present of different batches of a commercial chimeric

IgG1 mAb. Initial evaluations focused on the quantitative performance of the 12/13C6 2-AA labelling using N-glycan standards (oligomannose and bi-antennary N-glycans with and without core fucose, antennary galactose and sialic acid residues) to assess linearity and accuracy of the quantitative response. The molar ratios of the light and heavy labelled